Understanding tectonic stress and rock strength in the Nankai Trough accretionary prism, offshore SW Japan

Open Access
Author:
Huffman, Katelyn Allison
Graduate Program:
Geosciences
Degree:
Doctor of Philosophy
Document Type:
Dissertation
Date of Defense:
September 02, 2015
Committee Members:
  • Demian Saffer, Dissertation Advisor
  • Chris Marone, Committee Member
  • Donald Myron Fisher, Committee Member
  • Derek Elsworth, Committee Member
  • Maureen Feineman, Committee Member
Keywords:
  • Nankai Trough
  • in situ stress
  • unconfined compressive strength
  • borehole breakout
  • IODP
  • poroelastic FEM
Abstract:
Understanding the orientation and magnitude of tectonic stress in active tectonic margins like subduction zones is important for understanding fault mechanics. In the Nankai Trough subduction zone, faults in the accretionary prism are thought to have historically slipped during or immediately following deep plate boundary earthquakes, often generating devastating tsunamis. I focus on quantifying stress at two locations of interest in the Nankai Trough accretionary prism, offshore Southwest Japan. I employ a method to constrain stress magnitude that combines observations of compressional borehole failure from logging-while-drilling resistivity-at-the-bit generated images (RAB) with estimates of rock strength and the relationship between tectonic stress and stress at the wall of a borehole. This approach is commonly applied in boreholes, and has recently been applied to boreholes in the Nankai Trough. Although other methods of estimating tress magnitude exist (such as leak off tests, anelastic strain recovery, and focal mechanism or fault slip inversion), using compressional borehole breakouts allows for a characterization of stress with depth using commonly measured data. I use the method to constrain stress at Ocean Drilling Program (ODP) Site 808 and Integrated Ocean Drilling Program (IODP) Site C0002. Site 808 penetrat¬es the frontal thrust offshore Cape Muroto in the toe of the prism in the rupture zone of the 1944 Nankaido earthquake, and Site C0002 penetrates the inner most, deformed portion of the accretionary prism ~30m from the trench above the megasplay fault that slipped coseismically in the 1946 Tonankai Earthquake (both magnitude 8.1+ with large, devastating tsunamis). At Site 808, I consider a range of parameters (assumed rock strength, friction coefficient, breakout width, and fluid pressure) in the method to constrain stress to explore uncertainty in stress magnitudes and discuss stress results in terms of the seismic cycle. I find a combination of increased fluid pressure and decreased friction along the frontal thrust or other weak faults could produce thrust-style failure, without the entire prism being at critical state failure, as other kinematic models of accretionary prism behavior during earthquakes imply. Rock strength is typically inferred using a failure criterion and unconfined compressive strength from empirical relations with P-wave velocity. I minimize uncertainty in rock strength by measuring rock strength in triaxial tests on Nankai core. I find strength of Nankai core is significantly less than empirical relations predict. I create a new empirical fit to our experiments and explore implications of this on stress magnitude estimates. I find using the new empirical fit can decrease stress predicted in the method by as much as 4 MPa at Site C0002. I constrain stress at Site C0002 using geophysical logging data from two adjacent boreholes drilled into the same sedimentary sequence with different drilling conditions in a forward model that predicts breakout width over a range of horizontal stresses (where SHmax is constrained by the ratio of stresses that would produce active faulting and Shmin is constrained from leak-off-tests) and rock strength. I then compare predicted breakout widths to observations of breakout widths from RAB images to determine the combination of stresses in the model that best match real world observations. This is the first published method to constrain both stress and strength simultaneously. I find that stress is in the normal regime and approaches the strike slip regime with depth and rock strength, in lower P-wave rock, is comparable to existing empirical relations of P-wave velocity and unconfined compressive strength (UCS) and laboratory strength experiments on Nankai core; However, in high P-wave velocity rock, our strength constraints are lower than indicated by empirical relations of P-wave and UCS, suggesting there should be more thought given to employing these relations in high P-wave sediments in the Nankai trough. I find that stress change needed to produce active thrust faulting at Site C0002 is on the order of 12 MPa. Finally, I explore uncertainty in rock behavior during compressional breakout formation using a finite element model (FEM) that predicts Biot poroelastic changes in fluid pressure in rock adjacent to the borehole upon its excavation and explore the effect this has on rock failure. I test a range of permeability and rock stiffness. I find that when rock stiffness and permeability are in the range of what exists at Nankai, pore fluid pressure increase +/- 45° from Shmin and can lead to weakening of wall rock and a wider compressional failure zone than what would exist at equilibrium conditions. In a case example at, we find this can lead to an overestimate of tectonic stress using compressional failures of ~2 MPa in the area of the borehole where fluid pressure increases. In areas around the borehole where pore fluid decreases (+/- 45° from SHmax), the wall rock can strengthen which suppresses tensile failure. The implications of this research is that there are many potential pitfalls in the method to constrain stress using borehole breakouts in Nankai Trough mudstone, mostly due to uncertainty in parameters such as strength and underlying assumptions regarding constitutive rock behavior. More laboratory measurement and/or models of rock properties and rock constitutive behavior is needed to ensure the method is accurately providing constraints on stress magnitude.